This request for a competing continuation of DA011301-06, Vulnerability to drug abuse, responds to PA-04-100 Epidemiology of Drug Abuse. The application is based on the Great Smoky Mountains Study, a longitudinal, community-based study of 1,420 participants who assessed first in 1993, at age 9-13. They are now in their early 20's. This relatively poor, rural sample (25% of them American Indian) has a high prevalence of substance abuse or dependence (SUD) in adolescence, but there is evidence that this is leveling off as they reach adulthood. This final phase of GSMS will enable us to address several goals of the PA, including #4: "causal mechanisms leading to onset, maintenance, and remittance of drug abuse," and #5 "drug abuse over the life course, including developmental processes that influence drug use trajectories and behavioral, health, and social consequences of drug abuse." The focus of the application is to examine several biomarkers of stress in the context of stressor exposure: cortisol (baseline and reactivity) as a marker of acute stress, Dehydroepiandrosterone sulfate (DHEAS) as a marker of global well-being, C-reactive protein (CRP) as a measure of stress-induced inflammatory response, and Epstein-Barr Virus (EBV) antibodies as a marker of cell-mediated immune response. In addition, we examine PTSD symptoms as psychological responses to traumatic stressors. The goals of this application are: 1. To complete assays on already-collected blood spots, and to collect one more sample, to examine the relationships among biomarkers of stress, life stressors, and history of SUD. 2. To assess the sample at 25-26, using as before the Young Adult Psychiatric Assessment together with measures of personality, physical health, educational, work, and relationship history, and a wide range of risk and protective factors on which we have up to 8 previous assessments. 3. To compare those with persistent SUD with desisters and those who have not developed SUD on their history of stress (indicated by the specified biomarkers) in response to a wide range of putative stressors, to begin to identify causal mechanisms linking stressor exposure, stress, and SUD.